A Psychometric Comparison of Three Patient-Based Measures of Asthma Control

By Wallenstein, Gene V; Carranza-Rosenzweig, Jacqueline; Kosinski, Mark; Blaisdell-Gross, Bonnie; Et al

Key words: ACQ – ACT – Asthma – Asthma control questionnaire – Asthma control test Health-related quality of life – Measurement – Patient-reported outcomes – Rules of Two

ABSTRACT

Objective: Asthma is a multidimensional disease, characterized by changes in pulmonary function, transient and chronic symptoms, and effects on quality of life. In this study, we compared the psychometric properties and screening accuracy of three patient- based asthma control instruments including: the Asthma Control Test (ACT), a brief instrument developed to assess asthma control of patients in a clinical setting; the Asthma Control Questionnaire (ACQ), an instrument developed for use in clinical research; and the ‘Rules of Two’, a tool that has been used in both settings.

Methods: Patients (N= 313) completed the ACT, ACQ, and Rules of Two during two asthma clinic visits 4-12 weeks apart. Office staff recorded pre- and post-bronchodilator FEV1 measurements and asthma specialists provided a global rating of asthma control. Internal consistency reliability was computed and construct validity was evaluated using analysis of variance (ANOVA). Logistic regression and receiving operating characteristic (ROC) curve analysis was conducted to compare the screening accuracy of each measure in identifying patients with uncontrolled or moderate to severe asthma. The responsiveness of each measure to changes in asthma control and severity was tested using correlational and ANOVA methods.

Results: Results show that the ACT and ACQ have comparable reliability, validity, screening accuracy, and responsiveness. The Rules of Two, however, did not meet some standards and therefore has weaker psychometric properties.

Conclusion: The ACT and ACQ are comparable asthma control questionnaires. The choice of which questionnaire to use should be informed by considering several factors, such as the intended purpose and setting where the questionnaire will be used, as well as the content, practicality, availability of benchmark scores, and adaptability to multiple administration modes of each questionnaire. One potential limitation of the study is that the data were collected in a clinical setting with limited demographic information. Hence, additional studies are needed to evaluate the psychometric properties of each instrument across demographic and clinical subgroups of the general population.

Introduction

Approximately one out of every 13 individuals – adults and children alike – exhibit asthma symptoms in the United States every year1. In 2004, the prevalence of asthma in the United States was estimated at 8.1%1. Although the prevalence of asthma has increased over the last 20 years2″5, it seems to have recently stabilized6. The condition imposes a substantial burden on patients who commonly report limitations in their physical, emotional, and social functioning7-9, which are often exacerbated when symptoms are not well-controlled7’9. In terms of economic impact, the direct and indirect costs of asthma were approximated to be in excess of $16 billion in 20046.

Research over the last decade has contributed to an improved understanding of the pathophysiology of asthma and to the development of effective medications that help prevent acute attacks2. As a result, clinicians now focus on helping their patients achieve asthma control, which is one of the key components of the National Asthma Education and Prevention Program (NAEPP) 1997 Expert Panel Report: Guidelines for the Diagnosis and Management of Asthma and the update in 20022. Although these treatment guidelines are broadly recognized and endorsed by healthcare providers, studies have shown marked shortcomings in their implementation10″14, and that asthma control may be overestimated by both physicians and patients15.

The prevalence and costly burden of this condition highlight the need for a simple method to identify patients with poor asthma control. Assessing asthma control is complicated by the multidimensional nature of disease symptoms that are often modulated by a host of environmental variables and comorbid conditions. As a result, tests of pulmonary function, which provide a point-in-time estimate of asthma control, often correlate poorly with other tools for assessing asthma symptoms over extended durations. Moreover, as the treatment of asthma patients has increasingly shifted to the primary care setting where the availability of spirometry is limited16, a practical instrument for identifying and monitoring uncontrolled asthma is needed which can be easily scored and interpreted and which does not depend on physiological measurements of lung function.

In this study, we compared the psychometric properties and screening accuracy of three patientbased asthma control instruments including: the Asthma Control Test (ACT17), a brief questionnaire developed to assess asthma control of patients in a clinical setting; the Asthma Control Questionnaire (ACQ1819), an instrument developed for use in clinical research; and the ‘Rules of Two’ (see Appendix), a tool that has been used in both settings.

Methods

Measures

Asthma Control Test (ACT)

The ACT is a five-item self-administered patient-based measure of asthma control that was originally developed for use in clinical care settings17. There is evidence to support the reliability and validity of the ACT17’20. In the ACT, patients are asked to report their asthma symptoms, use of rescue medications, how asthma has impacted their daily functioning and sleep, and how patients perceive their asthma control. Each item has a five-level response scale and a 4-week recall period. The ACT is scored by summing responses for each of the five items to produce a final score ranging from 5 (poor control) to 25 (good control). ACT scores of 19 or less (scores <20) indicate that a patient's asthma may not be well-controlled17. The ACT can also be administered by telephone or face-to-face interview and can also be adapted for administration in almost any electronic mode, including interactive voice response (IVR), PDA, and desktop software21.

Asthma Control Questionnaire (ACQ)

The ACQ is a self-report patient-based asthma control questionnaire, validated for use with adults, that was originally developed for use in clinical research18. The ACQ contains seven items, six of which are completed by the patient. These six items ask patients to report on asthma symptoms, how asthma interferes with sleep/ awakening, limitations in activities due to asthma, and use of rescue medications. These items have a 1 -week recall period and are scored on a 7-point scale ranging from O (good control) to 6 (poor control). The last item consists of FEV1 percent predicted, as measured by clinical staff. The ACQ is scored by computing the mean of the response choice values across the seven ACQ items’8. Lower scores indicate better asthma control.

Rules of Two

The ‘Rules of Two’ is a self-report asthma control questionnaire consisting of three items which ask patients to report on asthma symptoms and rescue medication use (Appendix). Response options are dichotomous and are scored as 1 (yes) or 2 (no). Responses are summed and final scores range from 3 (poor control) to 6 (complete control).

Data collection

Patients

Patients (n = 313) who were visiting a physician from one of six asthma specialty groups for the first time were asked to participate in the study. The study was approved by Institutional Review Boards for each site. All patients and guardians signed an informed consent form. Patients were eligible for study participation if they were ≥ 12 years of age, had received a diagnosis of asthma, and were literate in English. Individuals were excluded from the study if they had other respiratory conditions or were involved in other clinical studies.

Procedures

After providing their informed consent, patients were administered (by a trained nurse or physician’s assistant) the ACT, ACQ, and Rules of Two instruments during two clinic visits – baseline and follow-up visits 4-12 weeks apart. After instrument completion at each visit, office staff recorded pre- and post- bronchodilator measurements of forced expiratory volume in 1 second (FEV^sub 1^). Asthma specialists, blinded to patient instrument responses, then provided a global rating of asthma control on a five- point scale ranging from ‘completely controlled’ to ‘not controlled at all.’ Asthma global control ratings were based on how well the NAEPP-defined goals of asthma therapy were being met as determined by the patient history, physical examination, and FEV^sub 1^. During the baseline visit, asthma specialists were also asked to indicate whether they intended to change the patient’s treatment course by ‘stepping down therapy’, ‘no change in therapy’ or ‘stepping up therapy”. This question was phrased in the past-tense in the follow- up assessment.

Assessments

Reliability

Internal consistency reliability for each of the asthma control instruments was assessed by calculating Cronbach’s alpha at baseline and follow-up. Internal consistency reliability is based on the intercorrelations of items within a scale. High inter-item correlations indicate that the items in a scale are all measuring the same construct22.

Validity

Construct validity of the three asthma control instrument\s was evaluated using the logic of ‘known groups’ validity23. For these validity tests, patients were categorized into groups known to differ in asthma control or severity derived from three clinical criterion measures: (1) specialists’ global rating of asthma control, (2) percent predicted FEV^sub 1^ values, and (3) change in therapy. The first criterion measure was the specialists’ global rating of asthma control, where patients were categorized into one of five groups based on this rating (ranging from ‘completely controlled’ to ‘not controlled at all’). The second criterion measure consisted of percent predicted FEV^sub 1^ values, where patients were categorized into four groups based on these values: (1) <60%, (2) 60-80%, (3) 81-100%, and (4) >100%. This stratification of percent predicted FEV^sub 1^ values was roughly based on the NAEPP breakdown of asthma severity (severe persistent, moderate persistent, intermittent persistent, and mild intermittent asthma)3. The third criterion measure consisted of whether or not the specialist changed the patient’s therapy as a result of the office visit. Patients were categorized into three groups: (1) stepped down therapy, (2) no change in therapy, and (3) stepped up therapy.

Table 1. Pearson correlation between asthma control measures and lung function at baseline assessment (N = 309)

Analysis of variance methods were used to compare mean ACT, ACQ, and Rules of Two scores across patients that differed on each criterion measure of asthma control. In addition, relative validity (RV) analyses were conducted to determine which asthma control instrument best discriminated among groups of patients differing in asthma control. Relative validity coefficients were calculated by dividing the F-statistic of each asthma control instrument by the largest F-statistic observed among all three. The relative validity coefficient for each instrument indicates, in proportional terms, its empirical validity relative to the best instrument in the test24,25. Pearson correlations between each asthma control instrument and the three asthma control criteria were also conducted (Table 1).

Screening accuracy in identifying uncontrolled asthma

Logistic regression methods and receiver operating characteristic (ROC) curve analyses were conducted to evaluate and compare the screening accuracy of all three asthma control instruments in identifying patients with uncontrolled asthma and differing in asthma severity. First, the specialists’ global rating of asthma control was used as the dependent variable in the logistic and ROC curve analyses to test the accuracy of each instrument for identifying uncontrolled asthma. A patient’s asthma was classified as uncontrolled (value of 1) if the specialists’ global rating was ‘not controlled at all,’ ‘poorly controlled,’ or ‘somewhat controlled.’ A patient’s asthma was classified as controlled (value of O) if the specialist’s rating was ‘well controlled’ or ‘completely controlled,’ second, the percent predicted FEV^sub 1^ was used as the dependent variable in the logistic and ROC curve analyses to test the ability of each instrument to detect moderate to severe persistent asthma. Patients were categorized as having moderate to severe persistent asthma if they had percent predicted FEV^sub 1^ values of <80% and patients with percent predicted FEV^sub 1^ values > 80% were categorized as intermittent mild or mild asthma3.

Responsiveness

The responsiveness of each asthma control measure was tested using correlational and analysis of variance methods. First, changes in scores on each asthma control measure from baseline to follow-up visit were correlated (Pearson) with changes in the specialists’ rating of asthma control and changes in lung function as measured by the change in percent predicted FEV^sub 1^ values and change in FEV^sub 1^ values. Second, analysis of variance was used to compare mean changes in scores on each asthma control measure across groups of patients who differed in the level of change in physician global ratings and change in FEV^sub 1^ values.

Change in specialists’ global ratings was derived by subtracting the baseline specialists’ global rating from the follow-up rating. Because so few patients deteriorated on the physician global rating over time, it was necessary to collapse the categories of change indicative of worsening into one category. Change in FEV^sub 1^ values was derived by subtracting the baseline FEV^sub 1^ values from the follow-up FEV^sub 1^ values and dividing by the baseline FEV^sub 1^ values to produce an indicator of the percent improvement in FEV^sub 1^ values. Patients were categorized into two groups according to percent improvement in FEV^sub 1^ values. The first group consisted of patients who improved by 10% or more from baseline. The second group consisted of patients who did not improve by 10% or more from baseline. This stratification of patients was deemed a minimally important change by asthma specialists involved in the development and validation of ACT26.

Results

Sample

A total of 313 patients participated in the study with a mean age of 35 years (SD = 15.3) ranging from 12 to 84 years. Data from 309 patients were analyzed at baseline and from 249 patients at follow- up. Approximately 5.4% of the sample was between the ages 12 and 14 years. There were no sub-analyses conducted on different age groups.

Based on specialists’ global ratings of asthma control at baseline, 48% of patients had ‘well controlled’ or ‘completely controlled’ asthma, 28% of patients had ‘somewhat controlled’ asthma and 23% had ‘poorly controlled’ or ‘not controlled at all’ asthma.

Approximately 8% of patients had mean percent predicted FEV^sub 1^ values ranging from 30 to 60%. Nearly 72% of the sample had mean percent of predicted FEV^sub 1^ values greater than 80%. Specialists changed therapy in 52% of patients as a result of the first visit. As shown in Table 2, therapy was stepped up in 147 patients and stepped down in 13 patients.

Reliability

In the present study, the average standardized intraclass correlation coefficient, a measure of internal consistency reliability (Cronbach’s alpha), was 0.85 (0.86 at follow-up) for the ACT, 0.89 (0.88 at follow-up) for the ACQ, and 0.66 (0.64 at follow- up) for the Rules of Two at baseline.

Validity

Table 1 presents the correlation between each asthma control instrument and physician global assessment of control and the two measures of lung function (percent predicted FEV^sub 1^ and FEV^sub 1^ values). The correlations were all statistically significant and in the hypothesized direction. In general, the strongest correlations were found between the asthma control questionnaires and the physicians’ assessment of control (Table 1). Correlations of the ACT and ACQ scales with the physician global assessment of control, percent predicted FEV^sub 1^ and FEV^sub 1^ values were similar in magnitude, while the Rules of Two scale exhibited relatively weaker correlations with these clinical indicators (Table 1).

Relative validity (RV) tests that estimate how well each asthma control instrument discriminates between groups of patients known to differ in physician global assessment of asthma control, lung function (measured by percent predicted FEV^sub 1^), and change in therapy are presented in Table 2. Based on the physician global assessment rating, patients with ‘completely controlled’ or ‘well controlled” asthma had significantly higher scores on the ACT scale as compared to patients with ‘somewhat controlled’ or ‘poorly controlled’ or ‘not controlled at all’ asthma (F = 27.5, p< 0.001). Similar results were obtained for the Rules of Two instrument (F= 16.0, p < 0.01). For the ACQ instrument, patients with 'completely controlled' or 'well controlled' asthma had significantly lower scores (F = 29.3, p < 0.001) than patients with 'somewhat controlled' or 'poorly controlled' or 'not controlled at all' asthma. Hence all three instruments were able to discriminate between groups that differed in terms of the physicians' global assessment of asthma control.

Table 2. Comparison of mean (SD) asthma control measure scores at baseline. Table 2 presents the results of construct validity tests of each asthma control measure in discriminating between groups of patients known to differ in the physician’s global assessment of asthma, lung function as measured by predicted FEV^sub 1^, and change in therapy

Similar results were observed among patients differing in their level of lung function (measured by percent predicted FEV^sub 1^), but in this case the ACT instrument had the highest relative validity (F = 16.1, p < 0.001), followed by the ACQ (F = 11.9, p < 0.01), and the Rules of Two (F = 5.9, p < 0.05).

Based on change in therapy criterion, patients with ‘stepped down’ therapy had significantly lower scores on the ACT (F = 32.9, p < 0.001) and Rules of Two instruments (F = 14.6, p < 0.01) compared to patients who were not prescribed a change in asthma therapy (Table 2). For the ACQ scale, significantly higher scores were observed among patients with stepped down therapy compared to patients who were not prescribed a change in therapy (F = 32.9, p < 0.001).

Both the ACT and the ACQ instruments had high relative validity in discriminating between patients that differed in the physician global assessment of asthma control (RV = 0.94 and RV = 1.00, respectively). The Rules of Two scale had the smallest relative validity (RV = 0.55) of all the three scales in discriminating between patients that differed in the physician’s global assessment of asthma control. In terms of the lung function criterion measure, the ACT instrument had the highest relative validity (RV = 1.00) in discriminating between groups of patients followed by the ACQ (RV = 0.74) and the Rules of Two (RV = 0.37). Lastly, against the change in therapy criterion measure, the ACT and ACQ instruments were e\qually valid (RV = 1.00) in discriminating between patients, followed by the Rules of Two (RV = 0.44).

Screening accuracy in identifying uncontrolled asthma

Screening accuracy results, based on physicians’ global assessment of asthma control, are presented in Table 3. The area under the ROC curve for both ACT and ACQ instruments was similar (0.77 and 0.78, respectively), while the Rules of Two was relatively lower (0.68). The ACT and ACQ instruments were also comparable in correctly classifying individual patients (68.4 and 71%, respectively) followed by the Rules of Two (66.6%).

Table 3. Performance of asthma control measures in screening for uncontrolled asthma*

Table 4. Performance of asthma control measures in screening for moderate or severe persistent asthma*

Screening accuracy in identifying moderate or severe persistent asthma

Based on lung function, patients were classified as having moderate or severe persistent asthma if they had percent predicted FEV^sub 1^ values less than 81%. Under this criterion measure, the number of patients with moderate or severe persistent asthma was 160 (52%) and the number of patients with mild intermittent or mild asthma was 148 (48%). Results for screening accuracy of moderate or severe persistent asthma based on lung function are presented in Table 4. The ACT had the highest positive (60.5%) and negative predictive values (76.9%), the highest area under the ROC curve (0.74), and the most patients correctly classified (74.9%). By comparison, the ACQ showed similar specificity (94.1%), negative predictive value (74.6%), percent correctly classified (72.6%) and area under the ROC curve (0.71) as the ACT. The Rules of Two showed a greater sensitivity (32.1%) and similar percent negative predictive value (76.3%) compared to ACT, but the lowest specificity (83.3%), predictive positive value (42.2%), percent correctly classified (69.1%) and area under the ROC curve (0.64).

Responsiveness

Table 5 presents correlations between changes in each asthma control scores and changes in physicians’ global assessment ratings, percent predicted FEV^sub 1^ and FEV^sub 1^ values. Changes in the ACT and ACQ instrument scores were significantly correlated with changes in each criterion measure of asthma control. Correlations between changes in Rules of Two scores and changes in percent predicted FEV^sub 1^ and FEV^sub 1^ values did not reach statistical significance. Correlations between the change in asthma control instrument scores and changes in physicians’ global assessment and FEV^sub 1^ were comparable in magnitude for both the ACT and ACQ (Table 5).

Results for mean changes in scores for each asthma control instrument across groups of patients differing in their level of change on the physicians’ global rating of control are presented in Table 6. Results support the construct validity of each asthma control instrument in responding to changes in the physicians’ global ratings of asthma control. The differences in mean change scores for each asthma control instrument were in the expected direction across levels of change in the physicians’ global ratings of asthma control. These differences in mean change scores as a function of changes in physician assessment of asthma control were statistically significant for all instruments (ACT, RV = 1.00; ACQ RV = 0.91; Rules of Two RV = 0.37; p < 0.001).

Table 7 presents the mean change scores for each asthma control instrument between groups of patients who did and who did not meet the threshold of change in FEV^sub 1^ values for 10% improvement. Results support the construct validity of each asthma control instrument in distinguishing between these two groups of patients. The differences in mean change scores on each asthma control instrument between the two groups were in the expected direction and the differences were statistically significant (ACT RV = 0.90; ACQ RV = 1.00; Rules of Two RV = 0.30; p < 0.001).

Discussion

This is the first study to conduct a head-to-head comparison of the ACT, ACQ, and Rules of Two asthma control instruments. Results show that the ACT and ACQ have desirable and comparable reliability, validity, screening accuracy, and responsiveness. However, the Rules of Two did not meet some standards and therefore has some limitations.

Nunnally and Bernstein27 have suggested minimum reliability coefficients of 0.70 for group-level analyses and values of 0.90 for individual-level analyses. The ACT and the ACQ (0.85 and 0.89, respectively) both exceeded the reliability standard for group- level analyses and were close to the reliability standard for individual-level analyses, while the Rules of Two (0.66) did not meet either threshold. These data indicate that the ACT and ACQ met commonly accepted reliability standards that are applied in clinical and research settings. The low reliability of the Rules of Two is likely due to it having only three items, but it may also be that these items are measuring disparate concepts.

Table 5. Correlation between changes in asthma control measure scores and changes in physician control ratings and lung function (N = 249)

Table 6. Mean changes in asthma control measure scores across groups of patients differing in the level of change in the physician global rating of asthma control

Table 7. Mean changes in asthma control scale scores between groups of patients differing in the level of change in FEV^sub 1^ values*

The ACT and ACQ had comparable criterion and construct validity, while the Rules of Two was less valid. In terms of criterion validity, although the Rules of Two correlated significantly with the three criterion measures of asthma control, these correlations were always of a lesser magnitude (0.21-0.40) compared to the ACT and the ACQ (0.28-0.53). Likewise, while the Rules of Two successfully discriminated among groups of patients differing in asthma control, it was a less valid tool in doing so when compared to the ACT and the ACQ.

In screening for moderate or severe persistent asthma as defined by percent predicted FEV1 values < 80%, the Rules of Two correctly classified only 69.1 % of patients, compared to 72.6% and 74.9% for the ACQ and ACT, respectively. This indicates that the ACT and ACQ have comparable screening accuracy in identifying positive cases of moderate or severe persistent asthma.

Overall, these results suggest that the Rules of Two has weaker psychometric properties than the ACT and the ACQ. Most fundamentally, the Rules of Two does not have adequate reliability. When results demonstrating its lesser validity, lower accuracy, and lesser responsiveness are also taken into account, it is clear that the ACT and the ACQ are highly preferable to the Rules of Two.

Both the ACT and ACQ perform well in terms of psychometric properties. When determining whether to use the ACT or the ACQ, one should consider the intended purpose and setting where the tool will be used. The decision of which tool to use may be best informed by considering the content, practicality, response burden, availability of benchmark scores, and adaptability to multiple administration modes of each questionnaire.

From a content perspective, the ACT and the ACQ have some comparable items. Both surveys contain questions on activity limitations due to asthma (ACT #1 and ACQ #3), shortness of breath (ACT #2 and ACQ #4), impact of asthma symptoms on sleep (ACT #3 and ACQ #1), and rescue medication use (ACT #4 and ACQ #6). The instruments differ because the ACT asks respondents to rate their asthma control (ACT #5), while the ACQ has separate questions on wheezing (ACQ #5), severity of morning asthma symptoms (ACQ #1), and requires lung function testing (ACQ #7).

The ACQ was primarily designed for use in clinical research trials while the ACT was initially designed for clinical use with patients, and, hence, the instruments may have advantages in the area for which they were primarily intended to be used. For instance, clinicians may be interested in how patients rate their own asthma control, information that can be obtained only from the ACT. As Nathan and colleagues17 noted, an example of how this question may be especially useful to clinicians is towards identifying patients who overestimate their asthma control, but may actually need further asthma education based on responses to other ACT items.

One important limitation of the current study is that data were gathered in a clinical setting. Implementing strict inclusion/ exclusion criteria as is done in clinical studies often results in reduced variation within the sample when compared to populations where these criteria are relaxed. This effect can often decrease the ability to detect psychometric differences in instruments, and thus, may have had an impact on the present findings. A second potential limitation is the lack of data on other demographic and clinical variables to test the generalizability of results. The psychometric properties of each instrument should be evaluated across demographic and clinical subgroups of the population.

Although the ACT and the ACQ have not been traditionally normed, benchmark ACT scores have been produced from the ACT developmental and longitudinal studies21. Benchmark scores are available for both new and established patients under the care of an asthma specialist21. In addition, benchmarks based on categories of asthma control derived from the physician global assessment of asthma control are available. These benchmark scores are potentially useful to both clinicians and researchers in comparing patients’ ACT scores to those of similar patients.

Conclusion

The ACT and ACQ are comparable asthma control questionnaires. The choice of which questionnaire to use should be informed by considering several factors, such as the intended purpose and setting where the questionnaire will be used, as well as the content, practicality, availability of benchmark scores, and adaptabil\ity to multiple administration modes of each questionnaire.

Acknowledgment

This study was supported by a grant from GlaxoSmithKline. Mark Kosinski and Priti Jhingran were involved in the original development of the Asthma Control Test.

CrossRef links are available in the online published version of this paper: http://www.cmrojournal.com

Paper CMRO-3554_3, Accepted for publication: 19 December 2006

Published Online: 17 January 2007

doi: 10.1185/030079906X167426

References

1. Behavioral Risk Factor Surveillance Survey, 2004. National Center for Environmental Health, Centers for Disease Control, [cited 25 Oct 2006]. Available from: http://www.cdc.gov/ asthma/brfss/04/ brfssdata.htm

2. National Institutes of Health/National Heart, Lung and Blood Institute, National Asthma Education and Prevention Program Expert Panel. Clinical practice guidelines: expert panel report 2 – guidelines for the diagnosis and management of asthma. NIH publication 97-4051. Bethesda, MD: 1997

3. National Institutes of Health/National Heart, Lung and Blood Institute, National Asthma Education and Prevention Program Expert Panel. Guidelines for the diagnosis and management of asthma: update on selected topics 2002. NIH publication 025075. Bethesda, MD: 2003

4. Mannino DM, Homa DM, Akinbami LJ, et al. Surveillance for asthma: United States 1980-1999. MMWR 2002;51(SS-1): 1-14

5. Mannino DM, Homa DM, Pertowski CA, et al. Surveillance for asthma: United States, 1960-1995. MMWR 1998;47(SS-1): 1-28

6. Trends in Asthma Morbidity and Mortality, July 2006. Epidemiology & Statistics Unit, Research and Program Services, American Lung Association, [cited 25 Oct 2006]. Available from: http:// www.lungusa.org/site/pp.asp?c=dvLUK9OOE&b=33347

7. Global Strategy for Asthma Management and Prevention. Global Initiative for Asthma. National Institutes of Health, National Heart, Lung, and Blood Institute. Revised 2002

8. Juniper EF. Quality of life in adults and children with asthma and rhinitis. Allergy 1997;52:971-7

9. Bousquet J, Knani J, Dhivert H, et al. Quality of life in asthma. I. Internal consistency and validity of the SF-36 questionnaire. Am J Respir Crit Care Med 1994;149:371-5

10. Meng Y-Y, Leung K-M, Berkbigler D, et al. Compliance with US asthma management guidelines and specialty care: a regional variation or national concern? J Eval Clin Pract 1999; 5: 213-21

11. Legorreta AP, Christian-Herman J, O’Connor RD, et al. Compliance with national asthma management guidelines and specialty care: a health maintenance organization experience. Arch Intern Med 1998; 158:457-64

12. Picken HA, Greenfield S, Teres D, et al. Effect of local standards on the implementation of national guidelines for asthma: primary care agreement with national asthma guidelines. J Gen Intern Med 1998;13:659-63

13. Moy JN, Grant EN, Turner-Roan K, et al. Asthma care practices, perceptions, and beliefs of Chicago-area asthma specialists. Chicago Asthma Surveillance Initiative Project Team. Chest 1999; 116:154-628

14. Warman KL, Silver EJ, McCourt MP, et al. How does home management of asthma exacerbations by parents of innercity children differ from NHLBI guideline recommendations? Pediatrics 1999; 103:422-7

15. Fuhlbrigge AL, Adams RJ, Guilbert TW, et al. The burden of asthma in the United States: level and distribution are dependent on interpretation of the national asthma education and prevention program guidelines. Am J Respir Crit Care Med 2002; 166:1044-9

16. Petty TL. John Hutchinson’s mysterious machine revisited. Chest 2002;121(5 Suppl):219-23S

17. Nathan RA, Sorkness CA, Kosinski M, et al. Development of the Asthma Control Test: a survey for assessing asthma control. J Allergy Clin Immunol 2004;113:59-65

18. Juniper EF, O’Byrne PM, Guyatt GH, et al. Development and validation of a questionnaire to measure asthma control. Eur Respir J 1999; 14:902-7

19. Juniper EF, Svensson K, Mork AC, et al. Measurement properties and interpretation of three shortened versions of the Asthma Control Questionnaire. Respiratory Medicine 2005; 99:553-8

20. Schatz M, Sorkness CA, Li JT, et al. Asthma Control Test: Reliability, validity, and responsiveness in patients not previously followed by asthma specialists. J Allergy Clin Immunol 2006;117:549- 56

21. Kosinski M, Bayliss MS, Turner-Bowker DM, et al. Asthma Control Test: A User’s Guide. Lincoln, RI: QualityMetric Incorporated, 2004

22. DeVellis RF. Scale Development: Theory and Applications. London: Sage Publications, 1991

23. Kerlinger FN. Foundations of Behavioral Research. New York: Holt, Rinehart and Winston, 1973

24. Liang MH, Larson MG, Cullen KE, et al. Comparative measurement efficiency and sensitivity of five health status instruments for arthritis research. Arth Rheum 1985;28:542-7

25. McHorney CA, Ware JE, Raczek AE. The MOS-36 Item Short Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care 1993;31:247-63

26. Schatz M, Sorkness CA, Li JT, et al. Asthma Control Test: Reliability, validity, and responsiveness in patients not previously followed by asthma specialists. J Allergy Clin Immunol 2006;117:549- 56

27. Nunnally JC, Bernstein IH. Psychometric Theory. New York: McGraw-Hill, 1994

Gene V. Wallenstein(a), Jacqueline Carranza-Rosenzweig(b), Mark Kosinski(a), Bonnie Blaisdell-Gross(a), Kavita Gajria(a) and Priti Jhingran(b)

a QualityMetric Incorporated, Lincoln, RI, USA

b GlaxoSmithKline, Research Triangle Park, NC, USA

Address for correspondence: Dr Gene V. Wallenstein, Outcomes Insight Consulting Division,

QualityMetric Incorporated, 640 George Washington Highway, Suite 201, Lincoln, RI 02865, USA.

Tel.: +1 401 334 8800, Ext 243; Fax: +1 401 334 8001; email: gwallenstein@qualitymetric.com

Appendix: ‘Rules of Two’ Questionnaire

The questions below ask about your asthma over different time periods.

1. Do you take your quick-relief inhaler more than TWO times per week? Yes/No

2. Do you awaken at night with asthma more than TWO times per month? Yes/No

3. Do you refill your quick-relief inhaler more than TWO times per year? Yes/No

Developed by the Baylor-Martha Foster Lung Care Center at Baylor University Medical Center

Copyright Librapharm Feb 2007

(c) 2007 Current Medical Research and Opinion. Provided by ProQuest Information and Learning. All rights Reserved.